RESERVA ECOLOGICA de GUAPIACU (REGUA) ‐ Results of Mammal Survey (10th April to 7th June 2013)

Steve Morgan 8/6/13

Results of Mammal Survey (May and June 2013)

Contents

1 Introduction

2 The Wetlands and The Inner Trails

3 The Fragment

4 The Outer Trails

5 The Surrounding Farmland

6 Conclusions and Recommendations For Further Study

Annex 1 ‐ Equipment and How it was Used

Annex 2 ‐ List of Mammals Recorded

Annex 3 ‐ Camera Trap Log

Annex 4 ‐ Identification Notes

I Introduction

The objective of this survey was to investigate the density and diversity of mammals in REGUA, in particular in those newer areas of the forest which have been more recently re‐established. Historically, mammals in REGUA have been heavily persecuted by hunting and their available habitat has been greatly reduced by human land use (coffee and banana plantations) and degraded by selective logging. However, the efforts of the REGUA team, since the reserve was established, have reversed these depressing trends and large new areas of forest have been re‐planted and a wetland area has been created. Hunting has been reduced to a much lower level, (not least by recruiting the erstwhile hunters as guides and rangers). The question now was to what extent were mammals re‐ colonising the forest?

The survey was conducted between 10/4/13 and 8/5/13 and covered the wetland and the inner network of trails ‐ and, to a more limited extent, the outer trails to the north and east. Some attention was given to a small area to the east known as "The Fragment" and the surrounding farmland.

A four‐pronged approach was used:

 walking the trails looking for animals and/or signs of animals (such as tracks, scrapes, scats)  spot‐lighting with high powered torches at night  live trapping using baited humane traps (ie "catch and release")  camera‐trapping, often at baited sites.

In addition, a bat detector was used to locate insectivorous bats, (though identification to species level was usually impossible using only this technique). A full inventory of equipment is given in Annex 1.

Each of these methods yielded useful information, though camera‐trapping, especially at pre‐baited sites, produced the most spectacular results. After all, there is no better way to determine whether mammals are present at a particular site, (and if they are, what species they might be), than by keeping it under twenty‐four hour surveillance!

Given the limitations of the time available, the logistical problems of accessing some areas and the modest amount of equipment that it was possible to transport to the Reserve, the conclusions presented here are highly impressionistic and provisional. However, it is hoped that they offer at least a snapshot of what is happening and a guide to further survey work. After only eight weeks in the field we might not necessarily know all the answers but perhaps we now have a better idea what questions to ask?

Finally, acknowledgement must be given to the guides and rangers at REGUA, whose knowledge and field‐craft was exceptional. The bird guide, Adilei, was of enormous assistance in advising on sites for camera traps and in finding and pointing out species such as Orange‐spined Hairy Dwarf Porcupine and Brown ‐throated Three‐toed Sloth. And the generous and patient help of Raquel and Jorge in providing bait and lifts to the outer trails was invaluable.

The Wetland and The Inner Trails

2.1 The Wetland, Its Capybara, Bats And Aquatic Mammals

The wetland, consisting of three pools, and the surrounding woodland is all fairly new, (having been established around seven years ago), and was previously pasture. However, it has quickly become naturalised and already supports a wide diversity of birds, insects and butterflies.

The three wetland pools support at least three groups of Capybara, each numbering around 10‐16 animals. (On 11/4/13 I made a night walk of the Yellow Trail and connecting trails and disturbed three distinct groups, two of which were on Wetland 2, the other on Wetland 3). There were also other individuals, suggesting a total population of around 45‐50. Later investigations have corroborated this initial estimate.

Capybara resting on an island in Wetland 2.

By day, the Capybara rest in the water or on the islands on Wetlands 2 and 3, in particular the island close to the wooden bird hide. Odd individuals rest on Wetland 1. By night, most of the Capybara leave their daytime haunts and go to graze in nearby fields. On several occasions large numbers, (possibly two out the three groups) were seen grazing the football pitch outside the Volunteer Houses. A third group was heard (on 3/5/13) heading along the Overflow Stream en‐route to the rough pastures east of REGUA where they joined the cows to graze. This group was subsequently caught by camera trap on the nights of the 4th, 5th and 6th repeating the journey, suggesting that this adventure might be a more or less regular nightly occurrence. This group consisted of at least 14 animals. It was interesting to see that Capybara were prepared to travel some0 50 metres or more to find good grazing. The wetland appears to be suitable habitat for Water Rat. However, painstaking night‐time searches along the margins and several live trapping attempts failed to confirm their presence. It seems likely that the physical barriers of surrounding farmland have so far inhibited their colonisation of this habitat, though this species is notoriously hard to see and could easily have avoided detection.

There was a rumour that Brasilian Tapir had been seen in the wetland. Absolutely no evidence was found to support this story, (no tracks, no sightings, no other field signs). Sadly, it seems highly improbable that Tapir are present.

Along with Capybara, the other mammalian species to occupy the wetland in significant numbers is the Velvety Free‐tailed Mastiff Bat (Molossus molossus). This species is easy to identify visually, (to genus level), by its distinctively long and narrow wings, its long tail and fast, agile flight. For some while, it was unclear whether the bats being observed were Molossus molossus or its close relative Molossus rufus (the Black Mastiff Bat). However, a chance capture of one individual on 5/5/13 enabled a close examination and so a precise identification then became possible, (ID Note 1).

Velvety Free‐tailed Mastiff Bat (captured incidentally as it fell into a laundry bucket).

These bats roost in roofs of the Volunteer Houses and Courtyard Buildings (and probably in other nearby structures). At about 40 minutes before dusk they begin to emerge and circle the open areas around the football pitch to hawk for insects. As the light fades they then go off to the wetlands to feed on the hordes of midges and mosquitoes that appear in the early evening. By about one hour after dusk the entire population had usually left the vicinity of the roost. The population appears to exceed 100 individuals from a rough count at the wetland. Observation at the wetland suggested that a few individuals were significantly larger and bulkier than the majority. These bats, still having long tails and long, narrow wings, were subsequently identified as Black Mastiff Bats (Molossus rufus) which are known often to co‐roost with Molossus molossus.

2.2 Paca Hill eand Th Lodge

Very early in the survey, Adilei pointed out a good site for Nine‐banded Armadillo (Dasypus novemcinctus) on the Brown Trail at the 200 metre mark. At least one recently excavated burrow was visible together with a well worn network of "runways" through the leaf litter. A camera trap was installed on 11/4/13 and the site baited with bananas. Two days later the camera had caught the Armadillo, two Pacas (Agouti paca) and a South‐eastern Common Opossum (Didelphis aurita), Pacas are known to live singly or in monogamous pairs and so it was presumed that the two animals (distinguishable by the patterns of white spots on their flanks) were male and female.

The site was re‐baited every few days thereafter for three weeks with bananas, chopped manioc, avocado, diced apple, corn and other items of fruit. Results were consistently good: the male Paca (who became known as "Porky") appeared on a nightly basis (usually at around midnight) and the South‐eastern Common Oppossum re‐appeared on two occasions. A probable Grey Four‐eyed Opossum (Philander oppossum) appeared on 19/4/13 in an interesting stand‐off with the male Paca, both animals seemingly astonished to encounter each other and very wary of the other's presence. See ID Note 2. White‐tufted Marmosets (Callithrix jacchus jacchus) appeared diurnally on a number of occasions as did a Tegu Lizard (on 17/4/13). On 19/4/13 a feral domestic cat made a surprise entrance. The Armadillo made only two further re‐appearances after its debut on the first night of camera‐trapping.

However, the Armadillo's active workings were very evident and a burrow with at least three entrances was discovered only five metres from the baited site.

The multi‐entrance burrow on Paca Hill.

A fourth hole was located about eight metres away and may or not have been part of the same system. Another cluster of holes, probably forming a second system lay about thirty or forty metres away to the west of the power line, possibly the territory of another animal. In all, it is safe to say that Paca Hill holds at least two and possibly up to four or five Armadillos (based on the number of active workings found).

Walking the trail along Brown 0‐250 and along the connecting trail by the Overflow Stream produced visual sightings of Armadillo on 1/5/13 and 2/5/13. On 4/5/13 an animal was disturbed at Brown 100 which shot off through the leaf litter at speed, presumably also an Armadillo. Spotlighting along the road from the Courtyard to the Lodge produced nothing, though about five or six attempts were made. Another researcher claimed to have found a South‐eastern Common Opossum on this road on or around 5/5/13.

A number of evenings were spent at the baited site in surveillance, (1st and 2nd May and 4th June). These produced further records of Armadillo and at least one other unidentified animal (probably a Paca). Small Fruit Bats were common on these evenings. Further evenings were spent outside the Lodge where a variety of bat species were present, including nectivorous Long‐tongued Bats (genus/species unknown), fruit bats (taking fruit from the bird table) and small, insectivorous bats echo‐locating at 25 Khz to 35 Khz . On one occasion a fruit bat actually sat on my shoulder for a few moments! No large mammals were seen at the Lodge.

Live trapping produced poor results with only a single Black Rat (Rattus rattus) at Brown 200 on 14/4/13 and on 5/6/13 another Black Rat and a Large‐headed Rice Rat.

In conclusion, Paca Hill appears to be populated with several Nine‐banded Armadillos, two Pacas and several Opossums (which probably include much wider areas in their home ranges). There are good numbers of bats of at least three or four genuses (insectivorous, frugivorous and nectivorous). And there are a few non‐indigenous visitors such as Black Rat and Feral Domestic Cat. Other small mammals appear to be present only in very low numbers.

2.3 The Courtyard and The Volunteer Houses

A surprisingly large number of mammals live commensally with the human inhabitants of REGUA.

A group of White‐tufted Marmosets visit the garden outside the kitchen two or three times a week to feed on fruit on the bird table. The group numbers up to seven individuals and appears to be the same group that visit the baited site at Brown 200, (having been observed moving between the two locations on two occasions).

Around the Courtyard and the road outside live one or more Black‐footed Pygmy Rice Rats (Oligozormys nigripes). On 2/5/13 an individual was captured in a garbage container and closely scrutinised; see ID Note 3. The same individual (or others of the same species) were observed on several occasions around the Courtyard running alongside the wall or along the roof beams.

Black Rats were observed in the Courtyard area several times, for example on the 30th May. For a few days thereafter, a Feral Cat, (a different animal to that caught on the camera trap at Brown 200), appeared, presumably attracted by these rodents.

The roosting Mastiff Bats have already been mentioned but other bat species also roost in the buildings here, including Black Myotis (Myotis migricans). On 17/4/13 an individual was found clinging to the external kitchen wall and examined, (ID Note 4). And on 20/4/13 another researcher found, in his room, a Seba's Short‐tailed Fruit Bat, (ID .Note 5)

Finally, a mammal was found to be occupying the roof of Casa 1 during the latter part of May. No identification was possible but the most likely culprit was thought to be Opossum.

Left: Black‐footed Pygmy Rice Rat captured incidentally in a garbage container. Right: Black Myotis.

2.4 The Brown Trail From 250 to 700 (Re‐planted Area)

Beyond the suspension bridge, the Brown Trail runs alongside a small stream for 450 metres. To the west lies a fairly large area of re‐planted forest. The ground is moist along most of this way and two tiny drainage ditches cross the trail. At Brown 650, the trail rises steeply onto drier ground where the ground cover is thinner and where there is continuous leaf litter on the forest floor.

Night walks here produced nothing. Even bat numbers were low. Baiting with jam and diced apple failed to produce the expected response from Opossums and live trapping also drew a complete blank, (despite a three night campaign from 3/5/13 to 5/5/13). There is a well‐worn game trail from the re‐planted area to the stream at Brown 350, apparently the work of Capybara. Otherwise, I found no evidence of mammalian activity whatever.

2.5 Big Round Top

At Brown 700 the trail divides and forms a loop which girdles the hill I have termed "Big Round Top". The Brown Trail continues to the south of the hill while the Light Blue Trail loops around the northern side. Being on higher ground, the soil is much drier here. The ground cover is much less scrubby and there is a continuous carpet of leaf litter. On the northern side, many of the trees are covered in vines. The forest here is significantly older and more established, the habitat correspondingly richer and more complex.

There is evidence along both sides of Armadillo excavations and, unsurprisingly, I captured one individual on a camera trap at Light Blue 900 on 2/5/13. The habitat looks excellent here for Armadillo and Paca though, disappointingly, I failed to record the latter species.

On 29/4/13 I captured a pair of Crab‐eating Foxes (Cerdocyon thous) on a camera trap at Brown 850 as they patrolled the trail heading west. These were the same pair recorded at Brown 1780 on 22/4/13.

Baiting with jam and diced apple at Brown 750 failed to elicit a response from Opossum, suggesting that they may not regularly frequent the hill. (Opossums seemed to come readily to this bait where they were present and so a "no show" seemed to strongly indicate their absence). Guianan Squirrel (Sciurus aestuans) was recorded on four occasions: at Brown 850 on 27/4/13, at Light Blue 700 on 18/4/13, at Light Blue 1000 on 3/5/13 and at Brown 1150 on 4/5/13). The hill, for whatever reason, seemed to provide good habitat for this species which elsewhere was scarce.

In conclusion, Big Round Top seems to be populated mainly by Nine‐banded Armadillos and Guianan Squirrels with regular visits from Crab‐eating Fox.

2.6 Small Round Top

Small Round Top, a smaller hill to the west of Big Round Top, is girdled by a continuation of the Light Blue Trail (which provides a short cut back to the Yellow Trail). It offers a similar habitat to its larger brother next door and, indeed, there were Armadillo excavations in a number of places, (all quite old). However, no fresh excavations were found.

Live trapping on the 19th and 20th produced nothing and no evidence could be found of small rodents.

Time did not permit a more detailed investigation of this area though, from the limited evidence gained, it seems that mammals are scarce in this area.

2.7 The Brown Trail From Fox Hill to The Schincariol Trail

Heading east from its junction with the Yellow Trail, the Brown Trail passes alongside an area of well established secondary forest on what is called here "Fox Hill". The hillside and trail is quite wet along here, a factor which permitted easy detection of animal tracks.

Evidence of Crab‐eating Fox, Opossum and other smaller mammals (unidentified) was found between Brown 2000 and Brown 1700 and on the strength of this a camera trap was installed from 18/4/13 to 22/4/13. This produced a Crab‐eating Raccoon on 19/4/13 and a pair of Crab‐eating Foxes on 22/4/13. (The latter appeared to be the same pair that later turned up at Brown 850 on 29/4/13). Further along, at Brown 1510, the main trail joins a small "trilinha" and a camera trap was installed from 26/5/13 to 28/5/13. This produced one Crab‐eating Fox and one Paca. It appeared, (from careful examination of tail and body fur), that this Fox was a separate animal to the pair previously captured by the cameras. From here to Brown 1150, the forest has been re‐planted relatively recently and there are open areas with extensive grass patches and scrub. Apart from the tracks of one Crab‐eating Fox track (15/4/13) and of an Opossum (Didelphis sp) at the stream at Brown 1300 on 19/5/13, little evidence of mammal activity was found. The Sao Jose Trail itself was not explored.

At Brown 1600 a spur heads north towards the Schincariol Trail. Just after the junction with the Sao Jose Trail is a stream and, on reports of a recent Tayra sighting a site here was heavily baited with fruit, manioc and sardines for several days. A camera trap was installed for three nights (15/4/13 to 17/4/13) but no activity was recorded. The absence here of any tracks in the ground was also noted, suggesting a low density of mammals in this area.

Night walks along this trail produced nothing apart from modest numbers of bats, (mostly frugivorous species). The impression gained was that the Brown Trail here acts as a "highway" for patrolling carnivores like Crab‐eating Fox, Crab‐eating Raccoon and Tayra but holds few resident mammals.

2.8 The Farm Road

The "Farm Road" leaves the Yellow Trail at Yellow 1700 and heads westwards towards a small farm outside of REGUA. To the west is a hill with well‐established secondary forest while to the east is newly re‐planted forest together with some scrub, grass and open areas.

A variety of animal tracks were found on this road, suggesting that mammals might regularly be present. These included Crab‐eating Fox (found in four or five places), Opossum (found in two places), smaller mammals (unidentified) and domestic dog. On 19/4/13 a site about 500 metres west of the junction with the Yellow Trail was heavily baited with fruit, bananas, manioc and corn. A camera trap was also installed on 25/4/13. However, nothing was recorded and a week later the bait remained untouched. Because of the poor initial results and the risk of camera theft, no further attempt was made to record mammals here.

2.9 The Yellow Trail

The Yellow Trail circles the wetland and passes through extensive areas of re‐planted forest. Three hides offer good vantage points to scan the wetlands. One is a concrete construction located next to the Lodge, another an elevated wooden structure on the Yellow Trail on the opposite side of the wetland, the third (known as "Amanda's Hide") a simple ground‐level structure offering views across Wetland 2.

From the concrete hide, on 19/4/13, a Brown‐throated Three Toed Sloth was seen on a distant Cecropia tree, a location that subsequently produced further sightings of the same animal. South‐ eastern Common Opossum had, I was told, previously been sighted in the trees around this hide but, despite a thorough search, no sign of this species was found.

At Amanda's Hide, four selected trees were baited with jam and diced apple for three nights to try for Opossum. Results were immediate: three of the four baited branches were stripped bare of bait on each night, (the fourth remaining untouched throughout). On the evening of 25/4/13, with fresh bait in place, I sat to watch one of the branches and was rewarded, within thirty minutes, by the appearance of an Opossum. In the torchlight, I had good views for fifteen seconds or so before the animal retreated. Initially, I identified the animal as South‐eastern Common Opossum. However, the next morning, to my surprise, I found I had captured it in a nearby live mammal trap (intended for Pygmy Rice Rat). Closer inspection suggested that it might in fact be a juvenile White‐eared Opossum (Didelphis albiventris),w a ne species for REGUA. (See ID Note 6).

Juvenile South‐eastern Common Opossum, (initially suspected to be White‐eared Opossum).

However, subsequent discussion with biologists at UERJ confirmed that it was actually a juvenile South‐eastern Common Opossum (Didelphis aurita).

At the wooden hide, there had previously been sightings of the very rare Bare‐tailed Woolly Opossum (Caluromys philander). However, despite trying on two occasions to spotlight the surrounding treetops, I failed to see it there. I subsequently discovered that the Opossum had been nesting in a coconut shell suspended from the roof of the hide but that this shell had been removed.

Crab‐eating Foxes were patrolling the trail along by Yellow 1470 early on the morning of 18/5/13 after heavy rain. A pair, (possibly the "Brown Trail" pair filmed earlier), left clear tracks after a day's continuous rain had stopped at about 04.00.

Live trapping around the Yellow Trail produced two captures of Large‐headed Rice Rat (Hylaeamys macrocephalus), both adjacent to very thick ground cover. (See ID. Note 7) No other small animals were captured in the traps, despite many nights of trying. This low number of captures was extremely disappointing and, given the presumed "naivity" of the animals in never having been trapped before, pointed ominously to a lack of rodents and small mammals in the dre‐plante areas.

Night walks failed to find any arboreal rodents such as Tree Rat or Spiny Mouse, despite a prolonged effort in likely areas with vines and inter‐locking canopy. There were also no Guianan Squirrels around the Yellow Trail.

Tracks of Crab‐eating Fox early on 18/5/13, Yellow 1470.

A small spur leaves the Yellow Trail at 2150 and winds its way up Fox Mountain. Exploring this trail on 29/5/13 I found three very old Armadillo holes but no recent signs of activity.

The two captured Large‐headed Rice Rats

Finally, on 9/5/13, two White‐tufted Marmosets were observed at Yellow 2400. Both animals, presumably less well habituated than their counterparts around the Courtyard, disappeared quickly.

3 The Fragment

I visited The Fragment, a small and isolated patch of forest about two kilometres east of Regua, on 1/5/13. There is some good looking and well‐established forest here, though it is surrounded by banana plantations and roads.

Three Guianan Squirrels were found in dense viny scrub near the end of the main trail, (close to the disused house), giving their distinctive "chik‐chik‐chik‐chuk" alarm calls. Also, two White‐tufted Marmosets were present about half way along the trail. Distant calls, possibly of Brown Capuchin, (which have been seen here before), were heard.

There was some evidence of Armadillos (old holes) and some of the well‐drained slopes looked potentially good habitat for this species. The Chief Scientist at Regua, (Jorge Bizarro), has seen this species here fairly recently. There are other possibilities for which time did not permit an opportunity to explore. The house at the end of the trail, used now only as a storage shed, might hold roosting bats. There were holes in the steep earth banks along the trail, suggesting that small rodents might be present. And, it would be interesting to see if Opossums were present ‐ and, if so, which species.

In conclusion, The Fragment would certainly justify further study.

4 The Outer Trails

4.1 The 4x4 Trail

The 4x4 Trail proved to be the most productive area both in terms of numbers of animals and in diversity of species. Two sites were chosen for camera‐trapping, about four hundred metres apart mid‐way along the trail. The upper site was regularly baited with bananas, corn and other items.

It was discovered that the trail lay within the territory of two Pumas, (a female and a juvenile), which were sighted by one of the rangers, Rildo, on or around 22/4/13 at a point about half way along the trail. A tree was later found nearby, (by another ranger Adilei), with extensive scratching and scats.

The "scratching tree". Left: evidence of scratching. Right: scats of various dates.

The widely differing ages of the scats, (some a few days old, some very old), suggested strongly that the tree was an important and regular territorial marker for the resident female Puma. One scat appeared to contain Sloth hair.

Subsequently, tracks were found along the lower part of the trail, (21, 22, 26 and 27th May), indicating that the Puma was venturing as far south as the Log Bridge. Good videos of both the female and the juvenile were obtained on the camera trap at 23.06 on 23/5/13 at a baited site about half way along the trail. A poor quality video was recorded on the second camera trap at the "Scratching Tree" the previous morning at 03.53 of just the female. The last definate evidence of the Pumas being in the area was on the 27th when a very fresh scat was found, (obviously from earlier that morning). On the 30th May fresh tracks were found on the access road between the cemetery and the "Wobbly Bridge" heading south. The Puma had evidently left the forest, at least temporarily, most likely heading towards the reservoir. Both Pumas then disappeared. No further tracks or video records transpired and it was assumed that the pair had moved to another part of their territory. This, however, seemed to stimulate a considerable amount of activity from other carnivores, specifically Ocelot, Tayra, Crab‐eating Raccoon and Crab‐eating Fox.

The Ocelot, a male, was first recorded on the camera trap at the baited mid‐way site on 21/5/13 at around 09.17 heading south. He then re‐appeared at 22.35 on 27/5/13 at the "Scratching Tree" where he sniffed the Puma's scent carefully before mounting an adjacent branch and scratching the bark. He left at 22.50, heading off‐trail into the forest. Significantly, perhaps, the 27th was the last date on which the Pumas were known to be present in the immediate vicinity. On the 30th, he visited the "Scratching Tree" again. Once more he circled the spot, sniffing cautiously, before spraying the tree to leave his own territorial marker. He left the tree at 05.04 heading north and was picked up by the second camera trap at the baited site at 05.22 as he continued up the trail. That evening, at 18.35, he was back at the "Scratching Tree" to re‐scent the spot, leaving shortly afterwards at 18.37. He returned on 4/6/13 at 17.38 and sprayed briefly and on 7/6/13 at 06.35 when he defecated on the ebase of th tree.

The Tayra visited the "Scratching Tree" at 14.28 on the 27th, apparently wasting no time after the Pumas' departure. (His appearance was recorded on camera). He emerged from the forest on the eastern side of the trail, marched straight up to the tree and furiously rubbed his head and neck against the bark to leave his scent. He then departed into the forest on the opposite side. On 3/6/13 at 17.26 he returned in pouring rain and scented once more by vigorous neck rubbing.

In the early hours of 28/5/13, the camera recorded a Crab‐eating Raccoon visiting the tree. The animal approached nervously, sprayed the ground and then rubbed his head and neck against the trunk to leave his own scent. It left immediately thereafter.

The following night, at 18.09 on the 28th, three Crab‐eating Foxes, (a female and two juveniles), appeared at the "Scratching Tree". There was much nervous sniffing before eventually one of the juveniles sprayed and neck‐rubbed. All three Foxes seemed uneasy and were quickly on their way, heading south. The foxes returned, at 05.28, on 3/6/13 for a repeat performance.

Then, on 6/6/13, at 13.01, a lone Coati approached the tree to sniff and briefly scent‐mark.

By now, six different species of carnivore had scent‐marked at the tree, in a four day period, in an apparent competition to establish territorial rights. However, this action occurred only after the Pumas had left on the morning of the 27th.

The baited site also produced a variety of other mammals, including Coati, Guianan Squirrel, South‐ eastern Common Opossum and Orange‐spined Hairy Dwarf Porcupine. A group of four Coatis were regular visitors, as were at least one Opossum. Two Squirrels visited on the 25th. And, on 4/6/13 at 20.46, a single Paca came to feed on bananas. The Porcupine was spotted resting in a nearby tree at around 11.00 on 7/5/13. The other notable record at this site not already mentioned was two Tayra on 30/5/13 at 07.50.

Elsewhere on the 4x4 Trail, a troop of Brown Capuchins were well seen at the top end of the trail, (near Casa Aribal, at around 12.00 on 7/5/13. The group consisted of at least five individuals. A Brown‐throated Three‐toed Sloth was also seen on the 7th, lower on the trail. Nine‐banded Armadillo holes were visible at various points along the trail, though very few seemed recent or active. Most holes were old and cobwebbed. Some other, smaller, holes were evident possibly indicating the presence of rodents. No attempt was made at live trapping because of the logistical difficulties of transporting traps. A single Guianan Squirrel was seen on the morning of 7/6/13 about a third of the way along the trail.

4.2 The Green Trail

The Green Trail seems to hold a relatively low mammal population.

The lower part of the trail produced little. Red‐rumped Agouti were seen (by Adilei) at Green 600, on 23/4/13 and clear tracks of at least two animals were visible between Green 550 and 650. Previously, Red‐rumped Agouti tracks had been seen at the second stream (green 1200) on 12/4/13. No further evidence of those species was found. The bamboo stands at Green 800 to 950 were subjected to several night‐time searches for Southern Bamboo Rat but none were found. In the hope of finding Water Opossum, the streams at Green 0 (by Casa Christine), Green 350 and Green 1150 were camera trapped on the nights of 22/4/13 and 23/4/13 but nothing materialised, (although this species had reportedly been seen at odd times at the last of these three sites). One positive finding was the presence of at least a dozen Great Fruit‐eating Bats (Artibeus lituratus) on 21, 22 and 23/4/13 in the Cecropia tree outside Casa Pesquisa. These were absent on all subsequent visits.

The middle Green Trail, (1200 to the Waterfall at 2700), is more mature forest and seems to hold a greater number and a greater diversity of mammals. The dry slopes along Green 1750 to 2200 provide good habitat for Armadillo and a number of recent or active workings were found, notably at Green 1850, 2150 and 2200.

Left: dry slope at Green 2000 with carpet of leaf litter; Right: fresh Armadillo working at Green 2150.

These slopes evidently provide territories for a number of individual animals.

A Tayra was seen by a visiting birder on 5/5/13 at or around Green 2200 though no further sign of this animal was seen subsequently. And on 10/5/13, both Orange‐spined Hairy Dwarf Porcupine and Brown‐throated Three‐toed Sloth were seen by another birder, Tim Bradbury, and Adilei at various points along this section of the Green Trail. Other Sloths, (or possibly re0sightings of the same animal), were reported by other visitors subsequently.

There was not time to explore the upper Green Trail, though a camping expedition on 30/4/13 failed to produce a sighting of Kinkajou, (which had been seen here by rangers a number of times before).

4.2 The Red and Orange Trails

An exploratory walk along the Red Trail on 12/4/13 produced Puma scrapes at several locations between Red 2000 and the summit at Red 3200. Further visits found scrapes at Red 1100 (20/4/13), Red 2600 (24/5/13) and Red 800 (10/5/13). This indicated that a Puma was using this trail on a fairly regular basis and, indeed, Adelei recorded a male Puma on his camera trap at Red 2550 on or around 23/4/13. I installed a camera here on 24/4/13 but no Puma activity was recorded and I removed it on 10/5/13.

The early visits (12/4/13 and 20/4/13) had produced extensive evidence of Collared Peccary between Red 2000 and Red 3200, (digging along the trail). This ceased thereafter, which might account for the Puma's apparent disappearance for several weeks.

The camera trap at Red 2600 produced occasional visits from South‐eastern Common Opossum (apparently two different individuals) and a Paca.

Left: Puma scrape at Red 800 (16/5/13); right: Puma scrape at Red 2800 (12/4/13).

The walk on 16/5/13 also produced Brown Capuchin (Cebus apella) at Red 800, briefly seen at around 200 metres high in the canopy. The only other encounter with primates on this trail was on 12/4/13 when Brown Howler Monkeys (Alouatta fusca) were heard vocalising some distance to the west of Red 3000.

The Orange Trail connects the Red Trail to the upper Green Trail, leaving the Red Trail at the 1100 metre marker. After about 150 metres it crosses two rough, boulder‐strewn streams, both of which look good potential habitat for Water Opossum (Chironectes minimus). This area was explored briefly on 16/5/13 and, though no evidence of mammals was found, it looked interesting enough to warrant further investigation.

4.3 The Dark Green and Grey Trails

The Dark Green Trail runs for 1.3 kilometres along the side of a well‐drained hill where there is a good carpet of leaf litter. These slopes appear to be ideal Armadillo habitat and there are holes of varying vintages all the way from DG 150 to DG1200. At DG1000, on 16/5/13, there was an apparently active hole with fresh earth outside the entrance.

Fresh hole, possibly leading to an occupied burrow, Dark Green 1000. Note fresh earth excavation.

The Grey Trail offers similarly good Armadillo habitat and a recent working was found at Grey 100.

Unfortunately, time did not permit further exploration of these trails.

4.5 The Blue Trail

The Blue Trail leaves the Green Trail at Green 250 and ascends a steep hill before joining the end of the Red Trail at about Blue 1750. Thereafter it crests a summit and descends abruptly to Casa Aribal and the top end of the 4x4 Trail. The trail comprises about three kilometres of mainly good forest habitat. The early four or five hundred metres pass through newer forest, in which there is a lot of scrub, but beyond that point it climbs through more mature forest with taller trees.

The best habitat lies between the junction with the Red Trail and the Casa Aribal. Here, there are numerous Armadillo workings with a number of fresh excavations. At least one older hole appears to have been used by Paca and has a "plug" of leaves in place. On 21/5/13 a troop of Brown Capuchin were seen at Blue 2500

The junction at Blue 1750 looked like a good point at which to intercept a patrolling Puma (or other cat) since it offered three directions from which an animal might appear. This optimism was reinforced by the presence of an old Puma scrape, indicating the use of this route by that species. Therefore, a camera trap was installed from 16/5/13 to 21/5/13. Disappointingly, this produced only two videos of Armadillo, (probably the same animal).

The middle Blue Trail, Blue 1300 to 1750, climbs quite a steep slope in a series of "switchbacks". There is good evidence of Armadillo here, with numerous old holes and a few fresher workings.

The lower Blue Trail was walked on a number of occasions but produced only one mammal sighting: a Coati very close to the trail at Blue 1300 at about 11.00 on 11/5/13. The animal was foraging for food and spent some time ripping the bark off a fallen tree before ambling off down the steep slope and out of sight. Live trapping at Blue 120 on the 24th and 25th May produced nothing.

5 The Surrounding Farmland

The surrounding farmland was remarkably poor in mammals with only a few indications of activity. The road from Regua to Matumbo is disturbed by traffic until late at night but still contains evidence of mammal activity. A daytime walk on 16/4/13 produced fresh roadside tracks of Opossum (Didelphis Spp) and Crab‐eating Fox (Cerdocyon thous).

Opossum tracks near Matumbo.

A subsequent night walk produced nothing either on the road or in the fields. However, further roadside tracks of Opossum and Crab‐eating Fox were found on 24/5/13 as well as Capybara tracks near the entrance to Regua.

The access road from Regua to the Casa Pesquisa passes through mosaics of arable land, woodland and open pasture alongside the eastern edge of the Regua reserve. The road was travelled frequently by vehicle and on foot, both by day and night. The most interesting finding was the previously mentioned Puma tracks on 30/5/1. See below.

Left:Puma track on access road 30/5/13. Right: fresh Armadillo hole near Casa Pesquisa.

In addition, a fresh Armadillo hole was found about 100 metres before the Casa Pesquisa, the only evidence of this species along the access road I could find. And a small rodent was seen running across the road early on 21/4/13, possibly a Grass Mouse (Akadon sp). Otherwise, no other evidence of mammals was found outside the Regua reserve.

6 Summary of Conclusions and Recommendations For Further Study

6.1 Diversity

The diversity of species appears to be good. I recorded evidence of 29 species, (through direct observation, live trapping, camera trapping and hearing and seeing field signs. A total of 73 species had previously been recorded in Regua and bat surveys had found 33 species of bat. This diversity provides a good base from which the forest can continue to recover.

6.2 Density

However, many species appear to be present in fairly low numbers. There are "hotspots" (such as the 4x4Trail) which hold good numbers of mammals but, equally, there are "dead zones" (such as Brown 250‐750) which seem almost empty. Density seems to be very patchy and strongly related to the maturity of the forest. Older, more established areas generally hold greater numbers of mammals whereas newly re‐planted areas hold correspondingly fewer. Given that the habitat is richer and more complex in the older areas, this much is not surprising. The general picture ‐ good diversity but lower than expected density ‐ appears to apply to many parts of the Atlantic Forest according to the biologists at UERJ.

Some species seem to be present in numbers that correspond to what one might expect, others less so. The following table is a qualitative view of what I found.

Numbers level with or exceeding Numbers slightly below Numbers expectation expectation disappointing

Capybara Grey Four‐eyed Opossum Other Primates Nine‐banded Armadillo Coati Collared Peccary Paca Crab‐eating Raccoon Small cats South‐eastern Common Opossum Crab‐eating Fox Tapiti Brown‐throated Three‐toed Sloth Brown Capuchin Small rodents Bats Orange‐spined Hairy Dwarf Porcupine Red‐rumped Agouti Puma Guianan Squirrel Small Opossums White‐tufted Marmoset Buffy‐headed Marmoset Tayra Water Rat, Bamboo Rat

I managed to form a view of the likely populations of some species.

There are definately three Pumas in Regua, (a male, a female and a juvenile). Perhaps there is potentially room for one more female territory. Other cats are present; notably a large male Ocelot. Otherwise, tracks of small cats were found on only one occasion, (possibly Oncilla). It seems likely that the numbers of the smaller cats are very low.

The Capybara around the wetland seem to comprise three groups, (each of 10‐15 animals), supplemented by other individuals. The total number might be around 40‐50.

There are at least three Crab‐eating Foxes around the wetland, (a pair and a singleton). And there is a group of three more on the 4x4 Trail, (a female and two juveniles). No other Foxes were found, though some other pairs and/or family groups undoubtedly occur elsewhere in the forest.

6.3 Re‐colonisation

The quickest re‐colonisers seem to have been bats, which is hardly surprising given their ability to fly. Armadillos also seem to have re‐established themselves successfully, possibly because they multiply quickly and tolerate a variety of habitats (making it easier to cross intervening areas between forest fragments). South‐eastern Common Opossum also appear to have re‐colonised new areas of forest quickly.

Small rodents seem to have been slow to re‐appear. I was particularly surprised at the lack of Forest Rats (Delomys dorsalis), Spiny Rats (Proechimys sp) and Tree Rats (Nelomys sp). These may all be present in the deeper forest but they were not found around the wetland in the newer areas.

Primates are still restricted to remote areas of more mature forest and have not yet reached the re‐ planted areas. The exception is the White‐tufted Marmoset (Callithrix jacchus jacchus) which has enthusiastically colonised the wetland area and The Fragment. The closely related Black‐pencilled Marmoset (Callithrix jacchus pencillata) ‐ which historically should prevail in the Atlantic Forest ‐ seems completely absent, at least from the inner trails.

Paca were recorded at three locations and seem to be recovering well from historic hunting pressure. A pair and a juvenile were caught frequently on video at Brown 200 and the wide availability of Armadillo holes around large areas of Regua affords potentially good habitat.

Small arboreal species such as Mouse Opossums will possibly appear gradually as the newer forest matures and the canopy develops. This process will probably also favour Guianan Squirrels which like more complex forest habitat with vines and horizontal inter‐locking branches.

Predators such as Lesser Grison and small cats, (Oncilla, Jaguarundi, Margay, Ocelot), will probably be late re‐colonisers since they all depend on a strong base of prey species such as small rodents, (which it seems is taking time to develop). On the other hand, Tayra appear to be doing well. At least four individuals were found on the 4x4 Trail and others were reported on the Green and Schincariol Trails.

6.4 Animal Behaviour

With the possible exception of the habituated White‐tufted Marmosets, all mammal species were extremely shy and elusive, probably reflecting intensive past hunting pressure. Even species such as Crab‐eating Fox, (normally having quite a short flight distance), were extraordinarily wary. Hunting still continues, albeit at a much reduced level, and it may be some whilee befor animals become more approachable.

This wariness was reflected in some notionally diurnal species, such as Collared Peccary, being nocturnal. No daytime Peccary activity was found, though Adilei caught the species on his camera trap after dark on the upper Red Trail. The Pacas on Paca Hill were, at least initially, extremely nervous about approaching the baited area, not appearing until five or six hours after dark and then carefully circling the area for up to an hour before beginning to feed confidently. (By the first week of June, after nearly eight weeks of baiting, this caution had subsided somewhat).

The camera traps revealed that some species were more enthusiastic trail‐walkers than others. Puma, Ocelot, Crab‐eating Fox and Crab‐eating Raccoon were all, as might be expected, glad to use trails as "highways" through the forest. Others, such as Tayra, Coati and Opossum simply ploughed through the forest. To some extent, species such as Armadillo and Opossum used regular "runways" through the forest floor and one of the best sites on the 4x4 Trail was where one of these runways intersected the trail. The animals' caution did not extend to scent marking where an extraordinary sequence of events took place between 27/5/13 and 8/6/13. The female Puma and her juvenile disappeared on the 27th having strongly marked their favourite scenting tree. Thereafter, Tayra, Ocelot, Crab‐eating Fox and Crab‐eating Raccoon all visited the Pumas' tree and took turns to superimpose their own scents in a competitive orgy of territorial marking.

6.5 Further Research

This survey could only ever scratch the surface of an area the size of Regua and there is huge scope for further study of mammals.

Some areas, for example Lengruba and the upper Green Trail, were not visited at all; others, such as The Fragment or the Orange Trail only briefly. Moreover, with only two camera traps available, the number of sites that could be monitored and even then often only for short periods. Some interesting topics for future study might include:

 a more prolonged monitoring of the Red Trail to track the movement of the male Puma and to see if smaller cats like Oncilla are present and to search for Bamboo Rat in the bamboo stands along Red 3000  more extensive monitoring of the Blue, Grey and Dark Green Trails  exploration of the Sao Jose Trail and the Schincariol Trail  an exploration of the upper Green Trail for species such as Kinkajou and Muriqui  an exploration of Lengruba  live trapping for small mammals in the outer trails  further exploration of the streams alonge th Green, Red and Orange Trails to search for Water Opossum.

Future projects should consider the use of at least five camera traps to provide greater and more continuous coverage.

Annex 1 ‐ Equipment and How it Was Used

Live Traps

Three live traps were used. These were simple "drop door" traps, activated by the animal touching a sensitive metal plate inside the cage. Bait was usually corn, diced fruit and seeds ( cheap breakfast cereal!). A water container was provided inside the trap to ensure that a captured animal could survive in a healthy state until it could be processed and released.

A wide variety of habitat types were tried in order to test how small mammal density varied. At each site the traps were placed in such a way as to make entry as easy as possible. Small quantities of seeds were offered in the immediate surrounding area to entice animals towards the traps. No trap was ever left for more than two nights and in most cases traps were checked each morning.

Spotlighting

Three torches were used: a LED LENSER X21 for scanning wide and distant areas, a LED LENSER P12 for routine close to medium range scanning and a small, low‐powered head torch for creeping about without spooking wary animals.

The usual approach was to move slowly and quietly along the trails, stopping frequently to listen for movement. At this point only the head torch would be used and its beam would be directed away from where animals were expected. On hearing movement, the P12 would be directed, at the appropriate moment, towards the sounds detected in the hope of making visual contact. The extremely powerful X21 was used mainly for scanning distant areas.

Camera Traps

Two Bushnell "Trophy Cam" traps were used, mounted on bank sticks by a screw thread or lashed by a strap to a convenient tree. These cameras activate by means of an infra‐red beam, triggered by movement in the area in front of it. There is no flash, or indeed any indication at all that the camera has been triggered, and so any animal caught by the beam is completely unaware that it is being filmed.

In some instances the camera would be installed at a strategic point on a trail, (for example at a three‐way junction or at a site where fresh field signs were present), in the hope of catching a passing animal. In others, a site would be pre‐baited and then the camera installed if interest was shown in the free bait.

Bat Detector

A small heterodyne detector was used with frequency range of 10 Khz to 130 Khz.

Annex 2 ‐ List of Mammals Recorded

A total of 29 species were recorded, as follows:

Species Latin Name CT/V Notes

Grey Four‐eyed Opossum Philander opossum V/CT One‐time visitor to Site 1 South‐eastern Common Opossum Didelphis aurita CT Red and Brown Trails, 4x4 Trail.

White‐tufted Marmoset Callithrix jacchus jacchus V/CT Frequent around Office/kitchen. Brown Capuchin Cebus nigritus V 4x4 trail. Brown Howler Monkey Alouatta fusca H Red Trail 3000.

Brown‐throated Three‐toed Sloth Bradypus variegatus V Resting in tree seen from Lodge Hide.

Crab‐eating Raccoon Procyon cancrivorus CT Brown 1780 Crab‐eating Fox Cerdocyon thous CT Brown 1780 Coati Mundi Nasua nasua V/CT Blue 1300. Tayra Eira barbara CT 4x4 Trail. Ocelot Leopardus Pardalis CT 4x4 Trail Puma Puma concolor CT 4x4 trail

Nine‐banded Armadillo Dasypus novemcinctus V/CT Site 1. Orange‐spined Hairy Dwarf Coendu villosus V 4x4 trail Porcupine (spinosus)

Paca Agouti paca CT/V Nightly visitor to Site 1 Red‐rumped Agouti Dasyprocta agouti T Tracks on Green Trail. Guianan Squirrel Sciurus aestuans V B850. Large‐headed Rice Rat Hylaeamys macrocephalus V Yellow 1800 in LMT. Black‐footed Pygmy Rice Rat Oligoryzomys Nigripes V Courtyard. Black Rat Rattus rattus V Brown 200 in LMT. Capybara Hydrochaeris V/CT Wetland. hydrochaeris Collared Peccary Tayassu tajuca T Red Trail.

Black Myotis Myotis nigricans V On kitchen wall. Great Fruit‐eating Bat Artibeus lituratus V Outside Casa Pesquisa. Seba's Short‐tailed Fruit Bat Carollia perspicillata V Juan's captured specimen. Fringe‐lipped Bat Trachops cirrhosus V Schincariol roost site. Spear‐nosed Bat Phyllostomus hastatus V Schincariol roost site. Velvety Free‐tailed Mastiff Bat Molossuss molossus V Outside Casa 1, wetland. Captive Casa 2. Black Mastiff Bat Molossus rufus V Outside Casa 1

Total Recorded 29

Key

V = visual record

H = heard

CT = camera trap

Annex 3 ‐ Camera Trap Log

The camera traps were in‐situ, at various sites, for a total of approximately 2500 hours. the breakdown of results was as follows:

Site Hours No Species Others Species

Paca Hill 320 5 Paca, SE Common Opossum, White‐tufted Tegu Lizard, Marmoset, Nine‐banded Armadillo, Four‐eyed Feral Cat Grey Opossum,

Wetland 450 5 Paca, Crab‐eating Fox, Crab‐eating Raccoon, Nine‐ None (other) banded Armadillo, Capybara

Farm Road 20 0 None None

Red Trail 400 2 SE Common Opossum, Paca None

Blue Trail 220 1 Nine‐banded Armadillo None

Streams 90 0 None None (Green Trail)

4x4 Trail 1050 9 Puma, Ocelot, Coati, Crab‐eating Fox, Crab‐eating Guan, Dove Raccoon, Tayra, Paca, Guianan Squirrel, SE Common Opossum

Total 2500 14

Annex 4 ‐ Identification Notes

Note 1: Velvety Free‐tailed Mastiff Bat (Molossus molossus)

A specimen was captured on 5/5/13 exhibiting the following characteristics:

 long free tail extending 20 mm beyond the inter‐femoral membrane.  bare inter‐femoral membrane with no hairs visible  ears triangular in profile and naked  dog‐like face with no nose leaf  brownish‐black soft fur  fore‐arm 37mm; body 69mm excluding free tail. On this basis, the specimen was firmly identified as Molossus molossus.

Note 2: Grey Four‐eyed Oppossum (Philander oppossum)

On the night of 19/4/13, the camera trap at Brown 200 recorded three sequences of an Oppossum with the following characteristics:

Note 3: Black‐footed Pygmy Rice Rat (Oligoryzormys nigripes)

A specimen was captured incidentally on 2/5/13 with the following characteristics:

 body length of 98mm  tail length of 140mm  pointed muzzle  whiskers medium length not extending past ear  feet white; rear feet long and narrow  tail bare and very long relative to body  fur grey; underparts pale grey.

The specimen was therefore identified as Oligoryzormys nigripes.

Note 4: Black Myotis (Myotis nigricans)

A specimen was found on the exterior wall of the kitchen on the morning of 14/413 with the following characteristics:

 soft, black body fur  blunt, roughly triangular muzzle with no noseleaf  tail membrane extending beyond legs, curled under body at rest  body length 52mm  forearm length 33mm  tail extending to edge of inter‐femoral membrane but not beyond.

The specimen was therefore identifed as Myotis nigricans.

Note 5: Seba's Short‐tailed Fruit Bat (Carollia perspicillata)

A researcher had found a specimen in his room overnight on 19/4/13 exhibiting the following characteristics:

 short, narrow muzzle  pointed, pear‐shaped ears  fur long and soft  underparts light grey, slightly grizzled  wings and inter‐femoral membrane grey  nose lead long and narrow, reaching eye when flattened  triangular chin with central tubercle  very short tail, well within inter‐femoral membrane  inter‐femoral membrane shorter than legs. The specimen was therefore identified as Carollia perspicillata.

Note 6: South‐eastern Common Opossum (Didelphis aurita)

On the night of 25/4/13 an Oppossum was captured in a live trap displaying the following characteristics:

 grizzled grey body fur, not brownish as in Didelphis aurita  white ears, dark at base and broad and rounded at tip (more rounded tips than D. aurita)  distinctively black feet  distinctive black eye‐stripe  black mark down centre of forehead coming to sharp point at level of eyes  tail root less prominent than in Didelphis aurita  under‐fur white, not brownish.

An initial identification of White‐eared Opossum (Didelphis albiventris) was made, a conclusion shared by a number of other expert mammal‐watchers, (including Jon Hall, Phil Telfer, Regina Ribeiro and Fiona Reid). However, juvenile White‐eared Opossum is difficult to separate from juvenile South‐eastern Common Opossum (both of which have white ears) and after detailed discussion with biologists from UERJ it was finally concluded that the animal was, in fact, Didelphis aurita.